EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

WGAN-GP for Synthetic Retinal Image Generation: Enhancing Sensor-Based Medical Imaging for Classification Models

Héctor Anaya-Sánchez iD, Leopoldo Altamirano-Robles iD, Raquel Díaz-Hernández iD, Saúl Zapotecas-Martínez iD

Sensors · 2024

DOI: 10.3390/s25010167 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, 2024. Verfasst von Héctor Anaya-Sánchez und 3 weiteren Autor:innen. Beteiligte Einrichtung: Computer Science Department, Instituto Nacional de Astrofísica Óptica y Electrónica, Luis Enrrique Erro No. 1, Sta. María Tonantzintla, Puebla 72840,…. Hauptthemen laut MEDLINE: Diabetic Retinopathy, Retina, Image Processing, Computer-Assisted, Diagnostic Imaging. Die Arbeit wurde bisher 20-mal zitiert (Crossref) und nennt 36 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Council of Humanities, Sciences, Technologies, and Innovation of Mexico (CONAHCyT).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate synthetic image generation is crucial for addressing data scarcity challenges in medical image classification tasks, particularly in sensor-derived medical imaging. In this work, we propose a novel method using a Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) and nearest-neighbor interpolation to generate high-quality synthetic images for diabetic retinopathy classification. Our approach enhances training datasets by generating realistic retinal images that retain critical pathological features. We evaluated the method across multiple retinal image datasets, including Retinal-Lesions, Fine-Grained Annotated Diabetic Retinopathy (FGADR), Indian Diabetic Retinopathy Image Dataset (IDRiD), and the Kaggle Diabetic Retinopathy dataset. The proposed method outperformed traditional generative models, such as conditional GANs and PathoGAN, achieving the best performance on key metrics: a Fréchet Inception Distance (FID) of 15.21, a Mean Squared Error (MSE) of 0.002025, and a Structural Similarity Index (SSIM) of 0.89 in the Kaggle dataset. Additionally, expert evaluations revealed that only 56.66% of synthetic images could be distinguished from real ones, demonstrating the high fidelity and clinical relevance of the generated data. These results highlight the effectiveness of our approach in improving medical image classification by generating realistic and diverse synthetic datasets.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Diabetic RetinopathyRetinaImage Processing, Computer-AssistedDiagnostic ImagingHumansAlgorithmsNeural Networks, Computer

Stichwörter der Autor:innen

AIGANsgenerative AImachine learningretinal images

Autor:innen und Institutionen

  • Computer Science Department, Instituto Nacional de Astrofísica Óptica y Electrónica, Luis Enrrique Erro No. 1, Sta. María Tonantzintla, Puebla 72840, Mexico.
  • Optics Department, Instituto Nacional de Astrofísica Óptica y Electrónica, Luis Enrrique Erro No. 1, Sta. María Tonantzintla, Puebla 72840, Mexico.

ORCID-Profile: Héctor Anaya-Sánchez, Leopoldo Altamirano-Robles, Raquel Díaz-Hernández, Saúl Zapotecas-Martínez

Förderung

  • National Council of Humanities, Sciences, Technologies, and Innovation of Mexico (CONAHCyT)

Zitieren

APA

Anaya-Sánchez, H., Altamirano-Robles, L., Díaz-Hernández, R., & Zapotecas-Martínez, S. (2024). WGAN-GP for Synthetic Retinal Image Generation: Enhancing Sensor-Based Medical Imaging for Classification Models. Sensors. https://doi.org/10.3390/s25010167

Vancouver

Anaya-Sánchez H, Altamirano-Robles L, Díaz-Hernández R, Zapotecas-Martínez S. WGAN-GP for Synthetic Retinal Image Generation: Enhancing Sensor-Based Medical Imaging for Classification Models. Sensors. 2024. doi:10.3390/s25010167

BibTeX
@article{AnayaSnchez2024,
  author = {Anaya-Sánchez, Héctor and Altamirano-Robles, Leopoldo and Díaz-Hernández, Raquel and Zapotecas-Martínez, Saúl},
  title = {WGAN-GP for Synthetic Retinal Image Generation: Enhancing Sensor-Based Medical Imaging for Classification Models},
  journal = {Sensors},
  year = {2024},
  doi = {10.3390/s25010167}
}
Kostenlos starten →